Book Review — “The Thinking Machine” by Stephen Witt

I picked up The Thinking Machine with high expectations. Stephen Witt had already impressed me with How Music Got Free, one of my standout reads years ago. Add the fact that this new work won the FT and Schroders Business Book of the Year in 2025, and curiosity became inevitable.

At the center stands Jensen Huang, born in Taiwan, raised partly in Thailand, and ultimately shaped in the United States. Witt paints him not as a mythic genius but as a relentless operator who learned early that survival in Silicon Valley requires both charm and steel.

One early story captures the contingency that defines Nvidia’s origin. Nvidia was not even Jensen’s idea. Two engineers from Sun Microsystems wanted someone to run their startup after their idea had been ignored by their bosses, so they approached Huang to run their company. It is one of those quiet forks in history. Nvidia might never have existed.

The naming of the company is another gem. The founders initially used “NV” as shorthand for “new venture.” After a failed attempt to secure the name Nvision, they settled on Nvidia, drawn from the Latin invidia, meaning envy.

Witt excels when he shows Huang’s contradictions. Watch Huang in public and he projects warmth and approachability. Internally, he can be ferocious. One former employee described being humiliated in front of dozens of executives after a minor mistake. Huang calculated the man’s total career earnings and demanded it be refunded. The employee said he barely slept for weeks. And this is something employees say he does regularly and intentionally.

It is uncomfortable reading, and Witt does not soften it. Yet he also shows the logic behind Huang’s intensity. In one episode involving a VR headset, Huang berated engineers over a five hour rendering time. After the outburst, the team reduced it to ten seconds. As one executive observed, speed changes behavior. When iteration becomes instant, excellence becomes possible.

The book also highlights the personal sacrifices that shadow many tech success stories. Huang’s wife, a brilliant electrical engineer, left her career because the couple could not secure childcare. When asked about this later, Huang said Laurie always believed in him. It is a revealing moment, though I found myself wishing Witt had pressed her perspective directly.

Where Witt truly shines is in explaining Nvidia’s improbable luck. CUDA, the company’s parallel computing platform, was not built with artificial intelligence in mind. Machine learning barely registered on the roadmap. In fact, when an early researcher approached Nvidia about using CUDA for neural networks, the company did not respond. The industry bias against neural nets was that strong.

Then the world changed.

Witt describes this convergence with quiet awe. Nvidia had built the perfect engine for a revolution it did not foresee. Few corporate histories capture serendipity this well. The rise of AI begins to look less like destiny and more like a sequence of fortunate collisions.

The geopolitical context is equally sharp. Witt reminds readers that Nvidia’s dominance rests heavily on the manufacturing might of TSMC. Taiwan’s so called silicon shield is not theoretical. It is one of the most important strategic realities in modern technology. Geopolitics is such a cloud over everything Nvidia is doing that the thought is paralyzing so much so that Jensen (that’s what he likes everyone to call him) has banned the company Chief Risk Officer from even considering any mitigation strategies.

Still, the book is not uncritical. Witt raises serious concerns about the energy demands of AI infrastructure and the environmental cost of ever more powerful chips. He also treats the techno-optimist claim that AI will inevitably “take care of us” with warranted skepticism. Unlike Karen Hao’s “Empire of AI” which I criticized very much, Witt’s pessimism about AI were expressed without overlooking the gains. The explanations were lucid and taken from a lot of angles.

The final chapter is especially striking. During Witt’s last interview, Huang reportedly erupted in expletives when the author pressed him on AI’s risks. It is a deeply human ending. The titan of accelerated computing, suddenly irritated and defensive.

The Thinking Machine, far from having the awe of How Much Got Free, still succeeds. Witt shows a company that was bold, lucky, occasionally ruthless, and often misunderstood even by itself. He also delivers one of the clearest business narratives of the AI era, lucid without being simplistic.

By the final page, one conclusion feels unavoidable. Nvidia did not plan the future of artificial intelligence. It built the hardware, and the future ran straight into it.